3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000,
4 * 2001, 2002, 2004, 2005, 2006, 2007, 2008, 2012 by Larry Wall and others
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
11 #ifndef HANDY_H /* Guard against nested #inclusion */
14 #if !defined(__STDC__)
22 # define Null(type) ((type)NULL)
27 =for apidoc AmU||Nullch
28 Null character pointer. (No longer available when C<PERL_CORE> is defined.)
30 =for apidoc AmU||Nullsv
31 Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
36 # define Nullch Null(char*)
37 # define Nullfp Null(PerlIO*)
38 # define Nullsv Null(SV*)
50 /* The MUTABLE_*() macros cast pointers to the types shown, in such a way
51 * (compiler permitting) that casting away const-ness will give a warning;
55 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
56 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
59 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
60 # define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
62 # define MUTABLE_PTR(p) ((void *) (p))
65 #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
66 #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
67 #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
68 #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
69 #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
70 #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
79 /* bool is built-in for g++-2.6.3 and later, which might be used
80 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
81 be sure _G_config.h will be included before this file. _G_config.h
82 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
83 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
84 g++ can be identified by __GNUG__.
85 Andy Dougherty February 2000
87 #ifdef __GNUG__ /* GNU g++ has bool built-in */
93 /* The NeXT dynamic loader headers will not build with the bool macro
94 So declare them now to clear confusion.
96 #if defined(NeXT) || defined(__NeXT__)
99 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
103 # endif /* !HAS_BOOL */
104 #endif /* NeXT || __NeXT__ */
115 /* a simple (bool) cast may not do the right thing: if bool is defined
116 * as char for example, then the cast from int is implementation-defined
117 * (bool)!!(cbool) in a ternary triggers a bug in xlc on AIX
120 #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
122 /* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
123 * XXX Should really be a Configure probe, with HAS__FUNCTION__
124 * and FUNCTION__ as results.
125 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
126 #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
127 # define FUNCTION__ __func__
129 # if (defined(_MSC_VER) && _MSC_VER < 1300) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
130 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
131 # define FUNCTION__ ""
133 # define FUNCTION__ __FUNCTION__ /* Common extension. */
137 /* XXX A note on the perl source internal type system. The
138 original intent was that I32 be *exactly* 32 bits.
140 Currently, we only guarantee that I32 is *at least* 32 bits.
141 Specifically, if int is 64 bits, then so is I32. (This is the case
142 for the Cray.) This has the advantage of meshing nicely with
143 standard library calls (where we pass an I32 and the library is
144 expecting an int), but the disadvantage that an I32 is not 32 bits.
145 Andy Dougherty August 1996
147 There is no guarantee that there is *any* integral type with
148 exactly 32 bits. It is perfectly legal for a system to have
149 sizeof(short) == sizeof(int) == sizeof(long) == 8.
151 Similarly, there is no guarantee that I16 and U16 have exactly 16
154 For dealing with issues that may arise from various 32/64-bit
155 systems, we will ask Configure to check out
157 SHORTSIZE == sizeof(short)
158 INTSIZE == sizeof(int)
159 LONGSIZE == sizeof(long)
160 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
161 PTRSIZE == sizeof(void *)
162 DOUBLESIZE == sizeof(double)
163 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
167 #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
168 # include <inttypes.h>
169 # ifdef INT32_MIN_BROKEN
171 # define INT32_MIN (-2147483647-1)
173 # ifdef INT64_MIN_BROKEN
175 # define INT64_MIN (-9223372036854775807LL-1)
190 #endif /* PERL_CORE */
192 #if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
193 # if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
194 # define PeRl_INT64_C(c) CAT2(c,LL)
195 # define PeRl_UINT64_C(c) CAT2(c,ULL)
197 # if QUADKIND == QUAD_IS___INT64
198 # define PeRl_INT64_C(c) CAT2(c,I64)
199 # define PeRl_UINT64_C(c) CAT2(c,UI64)
201 # if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
202 # define PeRl_INT64_C(c) CAT2(c,L)
203 # define PeRl_UINT64_C(c) CAT2(c,UL)
205 # define PeRl_INT64_C(c) ((I64TYPE)(c))
206 # define PeRl_UINT64_C(c) ((U64TYPE)(c))
211 # define UINT64_C(c) PeRl_UINT64_C(c)
214 # define INT64_C(c) PeRl_INT64_C(c)
218 #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
220 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
221 Please search CHAR_MAX in perl.h for further details. */
222 #define U8_MAX UINT8_MAX
223 #define U8_MIN UINT8_MIN
225 #define I16_MAX INT16_MAX
226 #define I16_MIN INT16_MIN
227 #define U16_MAX UINT16_MAX
228 #define U16_MIN UINT16_MIN
230 #define I32_MAX INT32_MAX
231 #define I32_MIN INT32_MIN
232 #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
233 # define U32_MAX UINT32_MAX
235 # define U32_MAX 4294967295U
237 #define U32_MIN UINT32_MIN
241 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
242 Please search CHAR_MAX in perl.h for further details. */
243 #define U8_MAX PERL_UCHAR_MAX
244 #define U8_MIN PERL_UCHAR_MIN
246 #define I16_MAX PERL_SHORT_MAX
247 #define I16_MIN PERL_SHORT_MIN
248 #define U16_MAX PERL_USHORT_MAX
249 #define U16_MIN PERL_USHORT_MIN
252 # define I32_MAX PERL_INT_MAX
253 # define I32_MIN PERL_INT_MIN
254 # define U32_MAX PERL_UINT_MAX
255 # define U32_MIN PERL_UINT_MIN
257 # define I32_MAX PERL_LONG_MAX
258 # define I32_MIN PERL_LONG_MIN
259 # define U32_MAX PERL_ULONG_MAX
260 # define U32_MIN PERL_ULONG_MIN
265 /* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
266 #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
267 #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
268 #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
270 #define Ctl(ch) ((ch) & 037)
273 =head1 SV-Body Allocation
275 =for apidoc Ama|SV*|newSVpvs|const char* s
276 Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
278 =for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
279 Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
282 =for apidoc Ama|SV*|newSVpvs_share|const char* s
283 Like C<newSVpvn_share>, but takes a literal string instead of a string/length
284 pair and omits the hash parameter.
286 =for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
287 Like C<sv_catpvn_flags>, but takes a literal string instead of a
290 =for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
291 Like C<sv_catpvn_nomg>, but takes a literal string instead of a
294 =for apidoc Am|void|sv_catpvs|SV* sv|const char* s
295 Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
297 =for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
298 Like C<sv_catpvn_mg>, but takes a literal string instead of a
301 =for apidoc Am|void|sv_setpvs|SV* sv|const char* s
302 Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
304 =for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
305 Like C<sv_setpvn_mg>, but takes a literal string instead of a
308 =for apidoc Am|SV *|sv_setref_pvs|const char* s
309 Like C<sv_setref_pvn>, but takes a literal string instead of a
312 =head1 Memory Management
314 =for apidoc Ama|char*|savepvs|const char* s
315 Like C<savepvn>, but takes a literal string instead of a string/length pair.
317 =for apidoc Ama|char*|savesharedpvs|const char* s
318 A version of C<savepvs()> which allocates the duplicate string in memory
319 which is shared between threads.
323 =for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
324 Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
326 =head1 Hash Manipulation Functions
328 =for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
329 Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
331 =for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
332 Like C<hv_store>, but takes a literal string instead of a string/length pair
333 and omits the hash parameter.
335 =head1 Lexer interface
337 =for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
339 Like L</lex_stuff_pvn>, but takes a literal string instead of a
345 /* concatenating with "" ensures that only literal strings are accepted as
347 #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
349 /* note that STR_WITH_LEN() can't be used as argument to macros or functions
350 * that under some configurations might be macros, which means that it requires
351 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
354 /* STR_WITH_LEN() shortcuts */
355 #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
356 #define newSVpvs_flags(str,flags) \
357 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
358 #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
359 #define sv_catpvs_flags(sv, str, flags) \
360 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
361 #define sv_catpvs_nomg(sv, str) \
362 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
363 #define sv_catpvs(sv, str) \
364 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
365 #define sv_catpvs_mg(sv, str) \
366 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
367 #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
368 #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
369 #define sv_setref_pvs(rv, classname, str) \
370 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
371 #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
372 #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
373 #define gv_stashpvs(str, create) \
374 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
375 #define gv_fetchpvs(namebeg, add, sv_type) \
376 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
377 #define gv_fetchpvn(namebeg, len, add, sv_type) \
378 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
379 #define sv_catxmlpvs(dsv, str, utf8) \
380 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
381 #define hv_fetchs(hv,key,lval) \
382 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
383 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
384 : HV_FETCH_JUST_SV, NULL, 0))
386 #define hv_stores(hv,key,val) \
387 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
388 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
390 #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
392 #define get_cvs(str, flags) \
393 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
396 =head1 Miscellaneous Functions
398 =for apidoc Am|bool|strNE|char* s1|char* s2
399 Test two strings to see if they are different. Returns true or
402 =for apidoc Am|bool|strEQ|char* s1|char* s2
403 Test two strings to see if they are equal. Returns true or false.
405 =for apidoc Am|bool|strLT|char* s1|char* s2
406 Test two strings to see if the first, C<s1>, is less than the second,
407 C<s2>. Returns true or false.
409 =for apidoc Am|bool|strLE|char* s1|char* s2
410 Test two strings to see if the first, C<s1>, is less than or equal to the
411 second, C<s2>. Returns true or false.
413 =for apidoc Am|bool|strGT|char* s1|char* s2
414 Test two strings to see if the first, C<s1>, is greater than the second,
415 C<s2>. Returns true or false.
417 =for apidoc Am|bool|strGE|char* s1|char* s2
418 Test two strings to see if the first, C<s1>, is greater than or equal to
419 the second, C<s2>. Returns true or false.
421 =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
422 Test two strings to see if they are different. The C<len> parameter
423 indicates the number of bytes to compare. Returns true or false. (A
424 wrapper for C<strncmp>).
426 =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
427 Test two strings to see if they are equal. The C<len> parameter indicates
428 the number of bytes to compare. Returns true or false. (A wrapper for
434 #define strNE(s1,s2) (strcmp(s1,s2))
435 #define strEQ(s1,s2) (!strcmp(s1,s2))
436 #define strLT(s1,s2) (strcmp(s1,s2) < 0)
437 #define strLE(s1,s2) (strcmp(s1,s2) <= 0)
438 #define strGT(s1,s2) (strcmp(s1,s2) > 0)
439 #define strGE(s1,s2) (strcmp(s1,s2) >= 0)
440 #define strnNE(s1,s2,l) (strncmp(s1,s2,l))
441 #define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
444 # define memNE(s1,s2,l) (memcmp(s1,s2,l))
445 # define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
447 # define memNE(s1,s2,l) (bcmp(s1,s2,l))
448 # define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
451 #define memEQs(s1, l, s2) \
452 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
453 #define memNEs(s1, l, s2) !memEQs(s1, l, s2)
458 * Unfortunately, the introduction of locales means that we
459 * can't trust isupper(), etc. to tell the truth. And when
460 * it comes to /\w+/ with tainting enabled, we *must* be able
461 * to trust our character classes.
463 * Therefore, the default tests in the text of Perl will be
464 * independent of locale. Any code that wants to depend on
465 * the current locale will use the tests that begin with "lc".
468 #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
476 =head1 Character classes
477 This section is about functions (really macros) that classify characters
478 into types, such as punctuation versus alphabetic, etc. Most of these are
479 analogous to regular expression character classes. (See
480 L<perlrecharclass/POSIX Character Classes>.) There are several variants for
481 each class. (Not all macros have all variants; each item below lists the
482 ones valid for it.) None are affected by C<use bytes>, and only the ones
483 with C<LC> in the name are affected by the current locale.
485 The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
486 C<U8>) as input and returns a boolean as to whether or not the character
487 represented by that octet is in the named class based on platform, Unicode, and
488 Perl rules. If the input is a number that doesn't fit in an octet, FALSE is
491 Variant C<isFOO_A> (e.g., C<isALPHA_A()>) will return TRUE only if the input is
492 also in the ASCII character set. For ASCII platforms, the base function with
493 no suffix and the one with the C<_A> suffix are identical. On EBCDIC
494 platforms, the C<_A> suffix function will not return true unless the specified
495 character also has an ASCII equivalent.
497 Variant C<isFOO_L1> operates on the full Latin1 character set. For EBCDIC
498 platforms, the base function with no suffix and the one with the C<_L1> suffix
499 are identical. For ASCII platforms, the C<_L1> suffix imposes the Latin-1
500 character set onto the platform. That is, the code points that are ASCII are
501 unaffected, since ASCII is a subset of Latin-1. But the non-ASCII code points
502 are treated as if they are Latin-1 characters. For example, C<isSPACE_L1()>
503 will return true when called with the code point 0xA0, which is the Latin-1
506 Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
507 point as input. If the code point is larger than 255, Unicode rules are used
508 to determine if it is in the character class. For example,
509 C<isWORDCHAR(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A WITH
510 MACRON in Unicode, and is a word character.
512 Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
513 (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
514 classification of just the first character in the string is tested.
516 Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but uses
517 the C library function that gives the named classification instead of
518 hard-coded rules. For example, C<isDIGIT_LC()> returns the result of calling
519 C<isdigit()>. This means that the result is based on the current locale, which
520 is what C<LC> in the name stands for. FALSE is always returned if the input
521 won't fit into an octet.
523 Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
524 returns the same as C<isFOO_LC> for input code points less than 256, and
525 returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
527 Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
528 (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
529 classification of just the first character in the string is tested.
531 =for apidoc Am|bool|isALPHA|char ch
532 Returns a boolean indicating whether the specified character is an
533 alphabetic character in the platform's native character set, analogous to
535 See the L<top of this section|/Character classes> for an explanation of variants
536 C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>
537 C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
539 =for apidoc Am|bool|isALPHANUMERIC|char ch
540 Returns a boolean indicating whether the specified character is a either an
541 alphabetic character or decimal digit in the platform's native character set,
542 analogous to C<m/[[:alnum:]]/>.
543 See the L<top of this section|/Character classes> for an explanation of variants
544 C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
545 C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC> C<isALPHANUMERIC_LC_uvchr>, and
546 C<isALPHANUMERIC_LC_utf8>.
548 =for apidoc Am|bool|isASCII|char ch
549 Returns a boolean indicating whether the specified character is one of the 128
550 characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
551 On non-ASCII platforms, it is if this
552 character corresponds to an ASCII character. Variants C<isASCII_A()> and
553 C<isASCII_L1()> are identical to C<isASCII()>.
554 See the L<top of this section|/Character classes> for an explanation of variants
555 C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
556 C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
557 library routine C<isascii()>. In these cases, the variants whose names contain
558 C<LC> are the same as the corresponding ones without.
560 =for apidoc Am|bool|isBLANK|char ch
561 Returns a boolean indicating whether the specified character is a
562 character considered to be a blank in the platform's native character set,
563 analogous to C<m/[[:blank:]]/>.
564 See the L<top of this section|/Character classes> for an explanation of variants
565 C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>
566 C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
567 platforms do not have the C library routine C<isblank()>. In these cases, the
568 variants whose names contain C<LC> are the same as the corresponding ones
571 =for apidoc Am|bool|isCNTRL|char ch
572 Returns a boolean indicating whether the specified character is a
573 control character in the platform's native character set,
574 analogous to C<m/[[:cntrl:]]/>.
575 See the L<top of this section|/Character classes> for an explanation of variants
576 C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>
577 C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>.
579 =for apidoc Am|bool|isDIGIT|char ch
580 Returns a boolean indicating whether the specified character is a
581 digit in the platform's native character set, analogous to C<m/[[:digit:]]/>.
582 Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
583 See the L<top of this section|/Character classes> for an explanation of variants
584 C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC> C<isDIGIT_LC_uvchr>, and
587 =for apidoc Am|bool|isGRAPH|char ch
588 Returns a boolean indicating whether the specified character is a
589 graphic character in the platform's native character set, analogous to
591 See the L<top of this section|/Character classes> for an explanation of variants
592 C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>
593 C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
595 =for apidoc Am|bool|isLOWER|char ch
596 Returns a boolean indicating whether the specified character is a
597 lowercase character in the platform's native character set, analogous to
599 See the L<top of this section|/Character classes> for an explanation of variants
600 C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>
601 C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
603 =for apidoc Am|bool|isOCTAL|char ch
604 Returns a boolean indicating whether the specified character is an
605 octal digit, [0-7] in the platform's native character set.
606 The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
609 =for apidoc Am|bool|isPUNCT|char ch
610 Returns a boolean indicating whether the specified character is a
611 punctuation character in the platform's native character set, analogous to
612 C<m/[[:punct:]]/>. Note that the definition of what is punctuation isn't as
613 straightforward as one might desire. See L<perlrecharclass/POSIX Character
614 Classes> for details.
615 See the L<top of this section|/Character classes> for an explanation of variants
616 C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>
617 C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
619 =for apidoc Am|bool|isSPACE|char ch
620 Returns a boolean indicating whether the specified character is a
621 whitespace character in the platform's native character set. This is analogous
622 to what C<m/\s/> and C<m/[[:space:]]/> match in a regular expression.
623 See the L<top of this section|/Character classes> for an explanation of variants
624 C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>
625 C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
627 =for apidoc Am|bool|isUPPER|char ch
628 Returns a boolean indicating whether the specified character is an
629 uppercase character in the platform's native character set, analogous to
631 See the L<top of this section|/Character classes> for an explanation of variants
632 C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>
633 C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
635 =for apidoc Am|bool|isPRINT|char ch
636 Returns a boolean indicating whether the specified character is a
637 printable character in the platform's native character set, analogous to
639 See the L<top of this section|/Character classes> for an explanation of variants
640 C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>
641 C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
643 =for apidoc Am|bool|isWORDCHAR|char ch
644 Returns a boolean indicating whether the specified character is a character
645 that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
646 in a regular expression. A word character is an alphabetic character, a
647 decimal digit, a connecting punctuation character (such as an underscore), or
648 a "mark" character that attaches to one of those (like some sort of accent).
649 C<isALNUM()> is a synonym provided for backward compatibility, even though a
650 word character includes more than the standard C language meaning of
652 See the L<top of this section|/Character classes> for an explanation of variants
653 C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
654 C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
656 =for apidoc Am|bool|isXDIGIT|char ch
657 Returns a boolean indicating whether the specified character is a hexadecimal
658 digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
659 and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
660 See the L<top of this section|/Character classes> for an explanation of variants
661 C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
664 =head1 Miscellaneous Functions
666 =for apidoc Am|U8|READ_XDIGIT|char str*
667 Returns the value of an ASCII-range hex digit and advances the string pointer.
668 Behaviour is only well defined when isXDIGIT(*str) is true.
670 =head1 Character case changing
672 =for apidoc Am|char|toUPPER|char ch
673 Converts the specified character to uppercase in the platform's native
674 character set, if possible; otherwise returns the input character itself.
676 =for apidoc Am|char|toLOWER|char ch
677 Converts the specified character to lowercase in the platform's native
678 character set, if possible; otherwise returns the input character itself.
682 Still undocumented are , PSXSPC, VERTSPACE, and IDFIRST, and the other
683 toUPPER etc functions
685 Note that these macros are repeated in Devel::PPPort, so should also be
686 patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
690 /* Specify the widest unsigned type on the platform. Use U64TYPE because U64
691 * is known only in the perl core, and this macro can be called from outside
694 # define WIDEST_UTYPE U64TYPE
696 # define WIDEST_UTYPE U32
699 /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
700 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
701 * bits of information are lost even on a 64-bit machine, but to get the
702 * compiler to optimize it out if possible. This is because Configure makes
703 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
704 * sizeof() guarantees that this evaluates to a constant true at compile time.
706 #define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
709 # define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128))
711 # define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
714 #define isASCII_A(c) isASCII(c)
715 #define isASCII_L1(c) isASCII(c)
717 /* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
718 * and the 8 possible permutations of those bits exactly comprise the 8 octal
720 #define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
722 /* ASCII range only */
723 #ifdef H_PERL /* If have access to perl.h, lookup in its table */
725 /* Character class numbers. For internal core Perl use only. The ones less
726 * than 32 are used in PL_charclass[] and the ones up through the one that
727 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
728 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
729 * their actual definitions are here. If that file has a name not used here,
732 * The first group of these is ordered in what I (khw) estimate to be the
733 * frequency of their use. This gives a slight edge to exiting a loop earlier
734 * (in reginclass() in regexec.c) */
735 # define _CC_WORDCHAR 0 /* \w and [:word:] */
736 # define _CC_DIGIT 1 /* \d and [:digit:] */
737 # define _CC_ALPHA 2 /* [:alpha:] */
738 # define _CC_LOWER 3 /* [:lower:] */
739 # define _CC_UPPER 4 /* [:upper:] */
740 # define _CC_PUNCT 5 /* [:punct:] */
741 # define _CC_PRINT 6 /* [:print:] */
742 # define _CC_ALPHANUMERIC 7 /* [:alnum:] */
743 # define _CC_GRAPH 8 /* [:graph:] */
745 #define _FIRST_NON_SWASH_CC 9
746 /* The character classes above are implemented with swashes. The second group
747 * (just below) contains the ones implemented without. These are also sorted
748 * in rough order of the frequency of their use, except that \v should be last,
749 * as it isn't a real Posix character class, and some (small) inefficiencies in
750 * regular expression handling would be introduced by putting it in the middle
751 * of those that are. Also, cntrl and ascii come after the others as it may be
752 * useful to group these which have no members that match above Latin1, (or
753 * above ASCII in the latter case) */
755 # define _CC_SPACE 9 /* \s */
756 # define _CC_BLANK 10 /* [:blank:] */
757 # define _CC_XDIGIT 11 /* [:xdigit:] */
758 # define _CC_PSXSPC 12 /* [:space:] */
759 # define _CC_CNTRL 13 /* [:cntrl:] */
760 # define _CC_ASCII 14 /* [:ascii:] */
761 # define _CC_VERTSPACE 15 /* \v */
763 # define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
765 /* The members of the third group below do not need to be coordinated with data
766 * structures in regcomp.[ch] and regexec.c */
767 # define _CC_IDFIRST 16
768 # define _CC_CHARNAME_CONT 17
769 # define _CC_NONLATIN1_FOLD 18
770 # define _CC_QUOTEMETA 19
771 # define _CC_NON_FINAL_FOLD 20
772 # define _CC_IS_IN_SOME_FOLD 21
774 * If more bits are needed, one could add a second word for non-64bit
775 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
776 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
777 * is used only for optimization (as of this writing), and differs in the
778 * Latin1 range from the ALPHA bit only in two relatively unimportant
779 * characters: the masculine and feminine ordinal indicators, so removing it
780 * would just cause /i regexes which match them to run less efficiently */
782 #if defined(PERL_CORE) || defined(PERL_EXT)
783 /* An enum version of the character class numbers, to help compilers
786 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
787 _CC_ENUM_ALPHA = _CC_ALPHA,
788 _CC_ENUM_DIGIT = _CC_DIGIT,
789 _CC_ENUM_GRAPH = _CC_GRAPH,
790 _CC_ENUM_LOWER = _CC_LOWER,
791 _CC_ENUM_PRINT = _CC_PRINT,
792 _CC_ENUM_PUNCT = _CC_PUNCT,
793 _CC_ENUM_UPPER = _CC_UPPER,
794 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
795 _CC_ENUM_SPACE = _CC_SPACE,
796 _CC_ENUM_BLANK = _CC_BLANK,
797 _CC_ENUM_XDIGIT = _CC_XDIGIT,
798 _CC_ENUM_CNTRL = _CC_CNTRL,
799 _CC_ENUM_PSXSPC = _CC_PSXSPC,
800 _CC_ENUM_ASCII = _CC_ASCII,
801 _CC_ENUM_VERTSPACE = _CC_VERTSPACE
802 } _char_class_number;
805 #define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
806 #define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
808 #if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
809 # if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
810 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
811 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8
812 #error Need to adjust order of swash_property_names[]
815 /* This is declared static in each of the few files that this is #defined for
816 * to keep them from being publicly accessible. Hence there is a small amount
819 static const char* const swash_property_names[] = {
832 #define PL_utf8_alnum PL_utf8_swash_ptrs[_CC_WORDCHAR]
833 #define PL_utf8_alnumc PL_utf8_swash_ptrs[_CC_ALPHANUMERIC]
834 #define PL_utf8_alpha PL_utf8_swash_ptrs[_CC_ALPHA]
835 #define PL_utf8_graph PL_utf8_swash_ptrs[_CC_GRAPH]
836 #define PL_utf8_digit PL_utf8_swash_ptrs[_CC_DIGIT]
837 #define PL_utf8_upper PL_utf8_swash_ptrs[_CC_UPPER]
838 #define PL_utf8_lower PL_utf8_swash_ptrs[_CC_LOWER]
839 #define PL_utf8_print PL_utf8_swash_ptrs[_CC_PRINT]
840 #define PL_utf8_punct PL_utf8_swash_ptrs[_CC_PUNCT]
843 EXTCONST U32 PL_charclass[] = {
844 # include "l1_char_class_tab.h"
847 # else /* ! DOINIT */
848 EXTCONST U32 PL_charclass[];
851 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
852 # define _CC_mask(classnum) (1U << (classnum))
853 # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
854 && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask(classnum)))
856 /* The mask for the _A versions of the macros; it just adds in the bit for
858 # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
860 /* The _A version makes sure that both the desired bit and the ASCII bit
862 # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
863 && ((PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask_A(classnum)) \
864 == _CC_mask_A(classnum)))
866 # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
867 # define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
868 # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
869 # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
870 # define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
871 # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
872 # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
873 # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
874 # define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
875 # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
876 # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
877 # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
878 # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
879 # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
880 # define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST))
882 /* Either participates in a fold with a character above 255, or is a
884 # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask(_CC_NONLATIN1_FOLD)))
886 # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
887 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
888 _generic_isCC(c, _CC_NON_FINAL_FOLD)
889 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
890 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
891 #else /* No perl.h. */
893 # define isALPHA_A(c) (isASCII(c) && isALPHA(c))
894 # define isALPHANUMERIC_A(c) (isASCII(c) && isALPHANUMERIC(c))
895 # define isBLANK_A(c) (isASCII(c) && isBLANK(c))
896 # define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
897 # define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
898 # define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
899 # define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
900 # define isLOWER_A(c) (isASCII(c) && isLOWER(c))
901 # define isPRINT_A(c) (isASCII(c) && isPRINT(c))
902 # define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
903 # define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
904 # define isSPACE_A(c) (isASCII(c) && isSPACE(c))
905 # define isUPPER_A(c) (isASCII(c) && isUPPER(c))
906 # define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
907 # define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
908 # else /* ASCII platform, no perl.h */
909 # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
910 # define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
911 # define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
912 # define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
913 # define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
914 # define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
915 # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
916 # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
917 # define isPRINT_A(c) (((c) >= 32 && (c) < 127))
918 # define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
919 # define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) \
920 || ((c) >= 58 && (c) <= 64) \
921 || ((c) >= 91 && (c) <= 96) \
922 || ((c) >= 123 && (c) <= 126))
923 # define isSPACE_A(c) ((c) == ' ' \
928 # define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
929 # define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
930 # define isXDIGIT_A(c) (isDIGIT_A(c) \
931 || ((c) >= 'a' && (c) <= 'f') \
932 || ((c) <= 'F' && (c) >= 'A'))
934 #endif /* ASCII range definitions */
936 /* Latin1 definitions */
938 # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
939 # define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
940 # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
942 /* continuation character for legal NAME in \N{NAME} */
943 # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
945 # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
946 # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
947 # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
948 # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
949 # define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
950 # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
951 # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
952 # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
953 # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
954 # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
955 #else /* No access to perl.h. Only a few provided here, just in case needed
956 * for backwards compatibility */
957 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
958 * >= AA test to speed up ASCII-only tests at the expense of the others */
959 # define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
960 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
961 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
962 || NATIVE_TO_UNI((U8) c) == 0xAA \
963 || NATIVE_TO_UNI((U8) c) == 0xB5 \
964 || NATIVE_TO_UNI((U8) c) == 0xBA)))
965 # define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
971 || NATIVE_TO_UNI((U8) c) == 0xA0)
974 /* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
975 * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
976 * therefore can use native functions. For those where C89 doesn't define a
977 * function, use our function, assuming that the EBCDIC code page is isomorphic
978 * with Latin1, which the three currently recognized by Perl are. Some libc's
979 * have an isblank(), but it's not guaranteed. */
981 # define isALPHA(c) isalpha(c)
982 # define isALPHANUMERIC(c) isalnum(c)
983 # define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
984 # define isCNTRL(c) iscntrl(c)
985 # define isDIGIT(c) isdigit(c)
986 # define isGRAPH(c) isgraph(c)
987 # define isIDFIRST(c) (isALPHA(c) || (c) == '_')
988 # define isLOWER(c) islower(c)
989 # define isPRINT(c) isprint(c)
990 # define isPSXSPC(c) isspace(c)
991 # define isPUNCT(c) ispunct(c)
992 # define isSPACE(c) (isPSXSPC(c) /* && (c) != '\v' (Experimentally making
993 these macros identical) */)
994 # define isUPPER(c) isupper(c)
995 # define isXDIGIT(c) isxdigit(c)
996 # define isWORDCHAR(c) (isalnum(c) || (c) == '_')
997 # define toLOWER(c) tolower(c)
998 # define toUPPER(c) toupper(c)
999 #else /* Not EBCDIC: ASCII-only matching */
1000 # define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1001 # define isALPHA(c) isALPHA_A(c)
1002 # define isBLANK(c) isBLANK_A(c)
1003 # define isCNTRL(c) isCNTRL_A(c)
1004 # define isDIGIT(c) isDIGIT_A(c)
1005 # define isGRAPH(c) isGRAPH_A(c)
1006 # define isIDFIRST(c) isIDFIRST_A(c)
1007 # define isLOWER(c) isLOWER_A(c)
1008 # define isPRINT(c) isPRINT_A(c)
1009 # define isPSXSPC(c) isPSXSPC_A(c)
1010 # define isPUNCT(c) isPUNCT_A(c)
1011 # define isSPACE(c) isSPACE_A(c)
1012 # define isUPPER(c) isUPPER_A(c)
1013 # define isWORDCHAR(c) isWORDCHAR_A(c)
1014 # define isXDIGIT(c) isXDIGIT_A(c)
1016 /* ASCII casing. These could also be written as
1017 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1018 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1019 which uses table lookup and mask instead of subtraction. (This would
1020 work because the _MOD does not apply in the ASCII range) */
1021 # define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
1022 # define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
1026 /* Use table lookup for speed; return error character for input
1028 #define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \
1029 ? UNI_TO_NATIVE(PL_latin1_lc[ \
1030 NATIVE_TO_UNI( (U8) (c)) ]) \
1031 : UNICODE_REPLACEMENT)
1032 /* Modified uc. Is correct uc except for three non-ascii chars which are
1033 * all mapped to one of them, and these need special handling; error
1034 * character for input out-of-range */
1035 #define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \
1036 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \
1037 NATIVE_TO_UNI( (U8) (c)) ]) \
1038 : UNICODE_REPLACEMENT)
1040 #ifdef USE_NEXT_CTYPE
1042 # define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c))
1043 # define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
1044 # define isASCII_LC(c) isASCII((unsigned int)(c))
1045 # define isBLANK_LC(c) isBLANK((unsigned int)(c))
1046 # define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
1047 # define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
1048 # define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
1049 # define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
1050 # define isLOWER_LC(c) NXIsLower((unsigned int)(c))
1051 # define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
1052 # define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
1053 # define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
1054 # define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
1055 # define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
1056 # define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
1057 # define toLOWER_LC(c) NXToLower((unsigned int)(c))
1058 # define toUPPER_LC(c) NXToUpper((unsigned int)(c))
1060 #else /* !USE_NEXT_CTYPE */
1062 # if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1064 /* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
1066 # define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
1067 # define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \
1068 && isalnum((unsigned char)(c)))
1070 # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
1072 # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
1075 # define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
1077 # define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
1079 # define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
1080 # define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
1081 # define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
1082 # define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1083 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1084 # define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
1085 # define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1086 # define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
1087 # define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1088 # define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1089 # define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1090 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
1091 # define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
1092 # define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? tolower((unsigned char)(c)) : (c))
1093 # define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? toupper((unsigned char)(c)) : (c))
1097 # define isALPHA_LC(c) (isascii(c) && isalpha(c))
1098 # define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
1099 # define isASCII_LC(c) isascii(c)
1101 # define isBLANK_LC(c) (isascii(c) && isblank(c))
1103 # define isBLANK_LC(c) isBLANK_A(c)
1105 # define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
1106 # define isDIGIT_LC(c) (isascii(c) && isdigit(c))
1107 # define isGRAPH_LC(c) (isascii(c) && isgraph(c))
1108 # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1109 # define isLOWER_LC(c) (isascii(c) && islower(c))
1110 # define isPRINT_LC(c) (isascii(c) && isprint(c))
1111 # define isPUNCT_LC(c) (isascii(c) && ispunct(c))
1112 # define isSPACE_LC(c) (isascii(c) && isspace(c))
1113 # define isUPPER_LC(c) (isascii(c) && isupper(c))
1114 # define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
1115 # define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
1116 # define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
1117 # define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
1120 #endif /* USE_NEXT_CTYPE */
1122 #define isPSXSPC_LC(c) isSPACE_LC(c)
1124 /* For internal core Perl use only. If the input is Latin1, use the Latin1
1125 * macro; otherwise use the function 'above_latin1'. Won't compile if 'c' isn't unsigned, as
1126 * won't match above_latin1 prototype. The macros do bounds checking, so have
1127 * duplicate checks here, so could create versions of the macros that don't,
1128 * but experiments show that gcc optimizes them out anyway. */
1130 /* Note that all ignore 'use bytes' */
1131 #define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1132 ? _generic_isCC(c, classnum) \
1134 #define _generic_swash_uni(classnum, c) ((c) < 256 \
1135 ? _generic_isCC(c, classnum) \
1136 : _is_uni_FOO(classnum, c))
1137 #define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1138 #define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
1139 #define isASCII_uni(c) isASCII(c)
1140 #define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1141 #define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
1142 #define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1143 #define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
1144 #define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
1145 #define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1146 #define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
1148 /* Posix and regular space are identical above Latin1 */
1149 #define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
1151 #define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
1152 #define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
1153 #define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
1154 #define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
1155 #define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
1156 #define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
1158 #define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
1159 #define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1160 #define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1161 #define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
1163 #define _generic_LC_uvchr(latin1, above_latin1, c) \
1164 (c < 256 ? latin1(c) : above_latin1(c))
1165 #define _generic_LC_swash_uvchr(latin1, classnum, c) \
1166 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1168 #define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1169 #define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1170 _CC_ALPHANUMERIC, c)
1171 #define isASCII_LC_uvchr(c) isASCII_LC(c)
1172 #define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
1173 #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
1174 #define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1175 #define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
1176 #define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1177 _is_uni_perl_idstart, c)
1178 #define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1179 #define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
1180 #define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1181 space under locale */
1182 #define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1183 #define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
1184 is_XPERLSPACE_cp_high, c)
1185 #define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1186 #define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1188 #define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
1191 #define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
1193 /* Everything whose name begins with an underscore is for internal core Perl
1196 /* If the input is in the Latin1 range, use
1197 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1198 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1199 * fact that ASCII characters have the same representation whether utf8 or not.
1200 * Note that it assumes that the utf8 has been validated, and ignores 'use
1202 #define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
1203 ? _generic_isCC(*(p), classnum) \
1204 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1206 TWO_BYTE_UTF8_TO_UNI(*(p), \
1210 /* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1212 #define _generic_func_utf8(classnum, above_latin1, p) \
1213 _generic_utf8(classnum, p, above_latin1(p))
1214 /* Like the above, but passes classnum to _isFOO_utf8(), instead of having a
1215 * 'above_latin1' parameter */
1216 #define _generic_swash_utf8(classnum, p) \
1217 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
1219 /* Like the above, but should be used only when it is known that there are no
1220 * characters in the range 128-255 which the class is TRUE for. Hence it can
1221 * skip the tests for this range. 'above_latin1' should include its arguments */
1222 #define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
1223 (UTF8_IS_INVARIANT(*(p)) \
1224 ? _generic_isCC(*(p), classnum) \
1225 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
1229 /* NOTE that some of these macros have very similar ones in regcharclass.h.
1230 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1231 * there, differing in name only by an underscore from the one here
1232 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1233 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1234 * points; the regcharclass.h ones are implemented as a series of
1235 * "if-else-if-else ..." */
1237 #define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1238 #define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
1239 #define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1240 utf8, the non-utf8 macro works
1242 #define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1243 #define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1244 #define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1245 _is_utf8_FOO(_CC_DIGIT, p))
1246 #define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
1247 #define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, is_utf8_xidcont, p)
1249 /* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1250 * IDFIRST is an alnum. See
1251 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
1252 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1253 * This used to be not the XID version, but we decided to go with the more
1254 * modern Unicode definition */
1255 #define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
1256 _is_utf8_perl_idstart, p)
1258 #define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1259 #define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
1261 /* Posix and regular space are identical above Latin1 */
1262 #define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
1264 #define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
1265 #define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
1266 #define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
1267 #define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
1268 #define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
1269 #define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1272 #define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1273 #define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1274 #define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
1276 /* For internal core Perl use only. If the input is in the Latin1 range, use
1277 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1278 * use the value given by the 'utf8' parameter. This relies on the fact that
1279 * ASCII characters have the same representation whether utf8 or not. Note
1280 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
1281 #define _generic_LC_utf8(macro, p, utf8) \
1282 (UTF8_IS_INVARIANT(*(p)) \
1284 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1285 ? macro(TWO_BYTE_UTF8_TO_UNI(*(p), *((p)+1))) \
1288 #define _generic_LC_swash_utf8(macro, classnum, p) \
1289 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1290 #define _generic_LC_func_utf8(macro, above_latin1, p) \
1291 _generic_LC_utf8(macro, p, above_latin1(p))
1293 #define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1294 _CC_ALPHANUMERIC, p)
1295 #define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
1296 #define isASCII_LC_utf8(p) isASCII_LC(*p)
1297 #define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1298 #define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1299 #define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1300 #define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
1301 #define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1302 #define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1303 #define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
1304 #define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1305 space under locale */
1306 #define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1307 #define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1308 #define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1309 #define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1311 #define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
1313 /* Macros for backwards compatibility and for completeness when the ASCII and
1314 * Latin1 values are identical */
1315 #define isALPHAU(c) isALPHA_L1(c)
1316 #define isDIGIT_L1(c) isDIGIT_A(c)
1317 #define isOCTAL(c) isOCTAL_A(c)
1318 #define isOCTAL_L1(c) isOCTAL_A(c)
1319 #define isXDIGIT_L1(c) isXDIGIT_A(c)
1320 #define isALNUM(c) isWORDCHAR(c)
1321 #define isALNUMU(c) isWORDCHAR_L1(c)
1322 #define isALNUM_LC(c) isWORDCHAR_LC(c)
1323 #define isALNUM_uni(c) isWORDCHAR_uni(c)
1324 #define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1325 #define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1326 #define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
1327 #define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1328 #define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1329 #define isALNUMC(c) isALPHANUMERIC(c)
1330 #define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1331 #define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1332 #define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1333 #define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1334 #define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
1336 /* This conversion works both ways, strangely enough. On EBCDIC platforms,
1337 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1338 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1339 * that is EOT on ASCII; ST on EBCDIC */
1340 # define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64)
1342 /* Line numbers are unsigned, 32 bits. */
1344 #define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
1346 /* Helpful alias for version prescan */
1347 #define is_LAX_VERSION(a,b) \
1348 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1350 #define is_STRICT_VERSION(a,b) \
1351 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1353 #define BADVERSION(a,b,c) \
1359 #define READ_XDIGIT(s) (isALPHA(*(s)) ? ((*(s)++ + 9) & 0xf) : (*(s)++ & 0xf))
1362 =head1 Memory Management
1364 =for apidoc Am|void|Newx|void* ptr|int nitems|type
1365 The XSUB-writer's interface to the C C<malloc> function.
1367 In 5.9.3, Newx() and friends replace the older New() API, and drops
1368 the first parameter, I<x>, a debug aid which allowed callers to identify
1369 themselves. This aid has been superseded by a new build option,
1370 PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
1371 there for use in XS modules supporting older perls.
1373 =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
1374 The XSUB-writer's interface to the C C<malloc> function, with
1375 cast. See also C<Newx>.
1377 =for apidoc Am|void|Newxz|void* ptr|int nitems|type
1378 The XSUB-writer's interface to the C C<malloc> function. The allocated
1379 memory is zeroed with C<memzero>. See also C<Newx>.
1381 =for apidoc Am|void|Renew|void* ptr|int nitems|type
1382 The XSUB-writer's interface to the C C<realloc> function.
1384 =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1385 The XSUB-writer's interface to the C C<realloc> function, with
1388 =for apidoc Am|void|Safefree|void* ptr
1389 The XSUB-writer's interface to the C C<free> function.
1391 =for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1392 The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
1393 source, C<dest> is the destination, C<nitems> is the number of items, and
1394 C<type> is the type. Can do overlapping moves. See also C<Copy>.
1396 =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1397 Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1400 =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1401 The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
1402 source, C<dest> is the destination, C<nitems> is the number of items, and
1403 C<type> is the type. May fail on overlapping copies. See also C<Move>.
1405 =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1407 Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1410 =for apidoc Am|void|Zero|void* dest|int nitems|type
1412 The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1413 destination, C<nitems> is the number of items, and C<type> is the type.
1415 =for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1417 Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1420 =for apidoc Am|void|StructCopy|type *src|type *dest|type
1421 This is an architecture-independent macro to copy one structure to another.
1423 =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1425 Fill up memory with a byte pattern (a byte repeated over and over
1426 again) that hopefully catches attempts to access uninitialized memory.
1428 =for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1430 PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1432 =for apidoc Am|void|PoisonFree|void* dest|int nitems|type
1434 PoisonWith(0xEF) for catching access to freed memory.
1436 =for apidoc Am|void|Poison|void* dest|int nitems|type
1438 PoisonWith(0xEF) for catching access to freed memory.
1442 /* Maintained for backwards-compatibility only. Use newSV() instead. */
1444 #define NEWSV(x,len) newSV(len)
1447 #define MEM_SIZE_MAX ((MEM_SIZE)~0)
1449 /* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1450 * overly eager compilers that will bleat about e.g.
1451 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
1452 #ifdef PERL_MALLOC_WRAP
1453 #define MEM_WRAP_CHECK(n,t) \
1454 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
1455 #define MEM_WRAP_CHECK_1(n,t,a) \
1456 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
1457 #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
1459 #define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (croak_memory_wrap(),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1462 #define MEM_WRAP_CHECK(n,t)
1463 #define MEM_WRAP_CHECK_1(n,t,a)
1464 #define MEM_WRAP_CHECK_2(n,t,a,b)
1465 #define MEM_WRAP_CHECK_(n,t)
1467 #define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1473 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
1474 * go through functions, which are handy for debugging breakpoints, but
1475 * which more importantly get the immediate calling environment (file and
1476 * line number, and C function name if available) passed in. This info can
1477 * then be used for logging the calls, for which one gets a sample
1478 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
1481 * - not all memory allocs get logged, only those
1482 * that go through Newx() and derivatives (while all
1483 * Safefrees do get logged)
1484 * - __FILE__ and __LINE__ do not work everywhere
1485 * - __func__ or __FUNCTION__ even less so
1486 * - I think more goes on after the perlio frees but
1487 * the thing is that STDERR gets closed (as do all
1488 * the file descriptors)
1489 * - no deeper calling stack than the caller of the Newx()
1490 * or the kind, but do I look like a C reflection/introspection
1492 * - the function prototypes for the logging functions
1493 * probably should maybe be somewhere else than handy.h
1494 * - one could consider inlining (macrofying) the logging
1495 * for speed, but I am too lazy
1496 * - one could imagine recording the allocations in a hash,
1497 * (keyed by the allocation address?), and maintain that
1498 * through reallocs and frees, but how to do that without
1499 * any News() happening...?
1500 * - lots of -Ddefines to get useful/controllable output
1501 * - lots of ENV reads
1504 PERL_EXPORT_C Malloc_t Perl_mem_log_alloc(const UV n, const UV typesize, const char *type_name, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname);
1506 PERL_EXPORT_C Malloc_t Perl_mem_log_realloc(const UV n, const UV typesize, const char *type_name, Malloc_t oldalloc, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname);
1508 PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
1511 # ifndef PERL_MEM_LOG_NOIMPL
1520 # if defined(PERL_IN_SV_C) /* those are only used in sv.c */
1521 void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1522 void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1529 #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1530 #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
1531 #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
1534 #ifndef MEM_LOG_ALLOC
1535 #define MEM_LOG_ALLOC(n,t,a) (a)
1537 #ifndef MEM_LOG_REALLOC
1538 #define MEM_LOG_REALLOC(n,t,v,a) (a)
1540 #ifndef MEM_LOG_FREE
1541 #define MEM_LOG_FREE(a) (a)
1544 #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1545 #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1546 #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
1549 /* pre 5.9.x compatibility */
1550 #define New(x,v,n,t) Newx(v,n,t)
1551 #define Newc(x,v,n,t,c) Newxc(v,n,t,c)
1552 #define Newz(x,v,n,t) Newxz(v,n,t)
1555 #define Renew(v,n,t) \
1556 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1557 #define Renewc(v,n,t,c) \
1558 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1561 #define Safefree(d) \
1562 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
1564 #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
1567 #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1568 #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1569 #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
1571 #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1572 #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1574 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
1576 /* Using bzero(), which returns void. */
1577 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
1580 #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1581 #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1582 #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1583 #define Poison(d,n,t) PoisonFree(d,n,t)
1585 #ifdef USE_STRUCT_COPY
1586 #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
1588 #define StructCopy(s,d,t) Copy(s,d,1,t)
1591 #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
1592 #define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
1596 # define Perl_va_copy(s, d) va_copy(d, s)
1598 # if defined(__va_copy)
1599 # define Perl_va_copy(s, d) __va_copy(d, s)
1601 # define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1606 /* convenience debug macros */
1608 #define pTHX_FORMAT "Perl interpreter: 0x%p"
1609 #define pTHX__FORMAT ", Perl interpreter: 0x%p"
1610 #define pTHX_VALUE_ (void *)my_perl,
1611 #define pTHX_VALUE (void *)my_perl
1612 #define pTHX__VALUE_ ,(void *)my_perl,
1613 #define pTHX__VALUE ,(void *)my_perl
1616 #define pTHX__FORMAT
1619 #define pTHX__VALUE_
1621 #endif /* USE_ITHREADS */
1623 /* Perl_deprecate was not part of the public API, and did not have a deprecate()
1624 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1625 CPAN::Unpack show any users outside the core. */
1627 # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
1630 #endif /* HANDY_H */
1634 * c-indentation-style: bsd
1636 * indent-tabs-mode: nil
1639 * ex: set ts=8 sts=4 sw=4 et: